open TELEMAC-MASCARET The mathematically superior suite of solvers

Thank you so much!! Setting the source region method worked :)

Hi,

So this time I am using a bit finer mesh and realistic manning's n values for the river (0.035 to 0.065) and for the rest of the catchment (0.04 to 0.4)(spatially varied roughness - as BOTTOM FRICTION in the geometry file).
The catchment is very steep and there are no depressions where the water can get stored. So, whatever water goes in should come out sooner or later. I also added some water as the base flow by using source regions.

The peaks of the resulting event hydrographs are becoming smaller with the increasing roughness but they don't get wider i.e. not showing the delay in response (which should be the ideal case with increased roughness). It means the volume of the water is decreasing which is not practically possible. So, the previous problem still persists. I am curious where is the water lost? and how can I get a delayed response?

Attached are the few input files and the screenshots of the observed (black) and the simulated hydrographs.

Best Regards!

Hi,

It is hard to have an idea without more details.
A couple of things I would check:
- Did you check the mass balance in the log/sortie listing?
- If you see strange things, test to run the case without the sources and without downstream boundary, this is easier to check mass balance
- Are you sure that your implementation of time and spatially varying rain is OK? Maybe try a simpler rain definition when making tests.
- Which method do you use to extract the hydrographs? Are the lines long enough so that you don't miss a part of the flow that could flow around them?

If you share the geometry and the listing I can try to have a look.

Kind regards
PL

Hi,

Yes, I checked the sortie file and mass balance. The volume in domain is way more than before for increased roughness it should go out. This case is same even if I use the point rainfall as input (hyetograph file).
To, extract the hydrograph, I integrate the flowrate along a cross-section (line is long enough). 'Gridded data to be integrated' > 'FLowrate along xy'

Attached are the geometry files for different bottom frictions.

Best Regards!

Hi,

Please ignore the previous 'Newfolder.zip' attachment.
Here is the updated geometry files (with 2 different sets of bottom friction) and the line across which I am extracting the hydrograph.

Hi,

I tested running your latest files (line_geo.zip) with a simplified set-up:
- no open boundaries
- CN = 90 (as defined in your files)
- OPTION FOR INITIAL ABSTRACTION RATIO = 2
- constant rainfall of 100 mm/hour during 1 hour (total rainfall of 100 m)

Two runs with your two mesh files using your BOTTOM FRICTION

The cas file I used is inspired from the test case pluie, see attached.

I get the following mass balance at the end of the simulations:
- in_sled_1.2eg_100_5_cn90_bfric.slf: FINAL VOLUME 741337.9 M3
- in_sled_1.2eg_100_5_cn90_bfric_17.slf: FINAL VOLUME 741334.7 M3
The difference between the two cases is 0.0004%.

I don't see any problem.

Are you sure about the method you use to extract your hydrographs or about your set-up?

Once more, I would not recommend to try having a hybrid between hydrological (higher parts of the catchment) and hydrodynamic models in this fashion, it's hard to garantee that the results in the coarser areas are ok.

Good luck,

Regards
PL

Can you please run the two scenarios of my model setup from the attachment? When I run the simulation , the final volume from these scenarios are 19353.77 M3 and 22022.94 M3.

I extracted the hydrograph the way as I mentioned in my previous post (a screenshot is included in the following attachment). Is there any other way too to extract the hydrograph?

Also, Can you please explain a bit more about why are the results from coarser mesh areas not reliable? I am only interested in the water volume from the those areas not the hydraulics. How does the mesh size affect the volume of water?

drive.google.com/file/d/1CPIiRP1Lw66NXwP...Da7/view?usp=sharing

Best Regards,
Nitesh

Hello,

I am very busy at the moment so I cannot test your model just now. Will try when I get more time.

Regarding hydrograph, I use either your method or the SECTIONS FILE method, which you also seem to use according to your files.

If the water inflows and rainfal are identical between the two runs, with only bottom friction changing from the mesh files, I would suggest you to run the simulations without any open boundary, that is no water is going to leave the domain. This way you should be able to check mass balance easily.

Regarding my remark on coarse mesh, my point is that you want to perform a purely hydrological runoff computation in a 2D model with very coarse elements. You need to think that for the flow generated from runoff will migrate downstream by means of solving the shalow water equations for a schematic flow case with very small water depths (max 1 mm?) on flat and large cells (500 m). The friction term will be large because of the very small water depth which can influence the actual propagation time because in reality, flow would be concentrated in ditches / talwegs with H > 0,1-0,5 m and with thus more reasonable friction terms. You actually cannot garantee that the propagation time is physical (unless you checked that of course) and the resuls are likely to be different from a analytical rainfall-runoff model.

It maybe works, but I wouldn't use a 2D hydrodynamic model outside of its reasonable range of application. If you don't have other choice than using a 2D model for the whole area, I would refine the cell size.

Kind regards
PL